Talking Robot
From: NASA Tech Briefs Insider - 09/19/2006

Researchers at Waseda University (Tokyo, Japan) have constructed a robotic
vocalization system that mimics the articulators used in human speech-making.
Called the WT Series, the system includes robotic equivalents of the vocal
cavity, tongue, vocal chords, cheeks, lips, teeth, soft palate, nasal cavity,
nose, and lungs. Constituent components of the vocal system were modeled from
soft plastics and polymers approximating the flexibility/rigidity of their
biological counterparts. Motors control jaw and lip movement. The WT Series
has articulated up to 50 phonemes found in the Japanese phonetic system, and
the robot is able to repeat sounds by means of speech recognition software. 

Scientists hope the current model will lead to the production of cellular
phones that can compress data by transmitting human vocal movement instead of
human voices. Researchers speculate the WT Series could lead to the
development of medical training devices for the vocally challenged and
learning devices for foreign language study. 

Read the entire article at:
Anthropomorphic Talking Robot - Waseda-Talker Series
http://www.takanishi.mech.waseda.ac.jp/research/voice/index.htm

Links:
Humanoid Robotics Institute at Waseda University
http://www.humanoid.waseda.ac.jp/

Takanishi Laboratory
http://www.takanishi.mech.waseda.ac.jp/index.htm

Robot research at Takanishi Laboratory
http://www.takanishi.mech.waseda.ac.jp/research/index.htm